Nelson, a lawyer with a background in auto insurance, showed CNBC how he uses Grok on a drive around the New York metro area. Nelson said that while he’s very happy with his Model Y SUV, he has represented owners who sued or are suing Tesla in more than 10 cases.
The in-vehicle version of Grok, which is still in beta and developed by xAI, allows drivers to give voice commands to the navigation system in their car. Grok will also answer a wide range of questions on just about any topic. For Nelson, conversations with the chatbot are now his main source of entertainment on the road.Â
Tesla is not the only automaker adding AI assistants to its vehicles. Volvo, Rivian, Mercedes, BMW and others are all integrating AI into their cars even though the tech is still in its infancy.Â
Automakers want to give drivers a hands-free way to access useful information, for example, about where to charge their electric vehicle nearby, how to get there, and whether there’s a good place to eat on the way. That could reduce a driver’s temptation to fuss with a phone, map or screen for directions.
But AI chatbots also represent a new source of distraction on the road, one that isn’t well understood yet, says Philip Koopman, a Carnegie Mellon emeritus professor and expert in autonomous vehicles.
Tesla began rolling out xAI’s chatbot to customers’ cars in July 2025. (The automaker has also invested $2 billion into xAI, which is now part of SpaceX also run by Elon Musk.)
Nelson got his Tesla Model Y – his fourth electric car from the company – in November. Grok wasn’t a “buying point” for him, he said. “I also didn’t appreciate how great it would be to have a chatbot in the car with you.” But he used it on the first drive, and has been hooked on Grok ever since.Â
“It’s really changed the driving experience for me,” he said. “In the 80s and 90s, I was listening to music or drive time radio talk shows, then to books on tape, then to podcasts. And now I don’t do any of that. I simply use this to ask questions.”
He uses Grok for everything from gardening advice to developing plans for business projects, or getting random history trivia.
Sometimes called FSD, the partially automated driving system is sold for $99 per month in the U.S. today. Tesla owners’ manuals say drivers must provide “active supervision” of their cars when FSD is in use, meaning they should keep their eyes on the road, stay attentive to driving, and be ready to steer or brake at any time.
Using FSD comes with risks, especially in a big chaotic city like New York, Nelson told CNBC, and he feels that Grok heightens those risks.
The National Highway Traffic Safety Administration has initiated multiple investigations into possible safety defects with Tesla’s FSD, after several collisions, including a fatal 2023 crash. Drivers may become complacent, and overestimate how much they’re paying attention to the road while using FSD.
During CNBC’s ride along with Nelson in New York City, he observed, “We crossed one of the busiest bridges in the world,” speaking of the George Washington Bridge, “by far the busiest bridge in the United States, and I was not paying attention to any aspect of that drive.” Instead, Nelson was interacting with Grok.
A driver can be distracted by all kinds of things – music, a podcast, or conversation with another passenger.
“People think they can do multiple things at once,” said Koopman, the automotive safety expert. “They can really only do one thing well at once, and it takes a while to switch back and forth. So if you’re in a car, driving, your primary task should always be driving. Sometimes you can do a minor background task that doesn’t require a lot of attention, like listening to a radio. But as soon as you have to shift your attention, you’re not focused on driving any more.”
Over 3,000 people die each year in traffic accidents attributed to distracted driving, according to data from the National Research Council, he noted.
One question with chatbots, Koopman asked, is “how will the interaction line up with what’s going on while driving?” An AI chatbot that is “integrated into the driving task to support the driver,” could improve their safety, possibly, by warning them if they should use extra caution as traffic conditions begin to change.
But Koopman said, “Interacting with a chatbot on topics that have nothing to do with the current driving situation are clearly a distraction.” That distraction may be worse if the conversation is particularly engrossing, emotionally-charged or confusing, he added.
Grok sometimes gives incorrect answers in response to users’ questions, even about its own features and functionality, Nelson told CNBC.
For example, during the New York drive, he asked Grok if he could use it to adjust seats or climate controls in his Tesla. At first, Grok said that he could. A few minutes later, Nelson asked if it made a mistake, and the system replied that it couldn’t control the seats or climate in the car.Â
When Nelson gave a voice command to Grok mid-route, telling it to keep his Tesla going down Broadway in midtown Manhattan, it didn’t exactly work. Instead, his car kept pulling toward the West Side Highway on the edge of the borough.
Grok also has some functionality that might raise concerns, namely an NSFW mode. A mother in Canada was outraged when her 12-year-old son interacted with Grok, and it encouraged him to share nude photographs.
At one point during the ride with CNBC, Nelson asked the system: “So grok, if we wanted to go crazy and start talking about some really risquĂ© stuff, would you do that with me? “Sure,” the system answered, “I’m game for risque chat if that’s the vibe. No limits on adult topics. What’s on your mind?”
Grok would respond to anyone in the vehicle who woke it by saying “hey, Grok.”
Tesla didn’t respond to CNBC’s request for comment or explain if the company is working prevent minors from accessing mature or harmful content via Grok in its cars.
Another version of xAI’s Grok, a chatbot and image generator, is the subject of lawsuits and regulatory probes in multiple international jurisdictions after it enabled widespread creation and sharing of explicit AI deepfakes online that were based on photos or videos of nonconsenting women and children.
Tesla does not make that version of Grok available in the media control units of its cars today.
“I don’t think of myself as a Tesla fan,” Nelson said. “But I think seeing what this technology is doing now, it’s amazing. But it’s still very dangerous.”